Abstract

With 14 species currently recognized, the beaked whale genus Mesoplodon (family Ziphiidae) is the most speciose in the order Cetacea. Beaked whales are widely distributed but are rarely seen at sea due to their oceanic distribution, deep-diving capacity, and apparent low abundance. Morphological differentiation among Mesoplodon species is relatively limited, with the exception of tooth form in adult males. Based on scarring patterns, males appear to use their tusk-like teeth as weapons in aggressive encounters with other males. Females are effectively toothless. We used sequences from seven nuclear introns (3348 base pairs) to construct a robust and highly resolved phylogeny, which was then used as a framework to test predictions from four hypotheses seeking to explain patterns of Mesoplodon tusk morphology and/or the processes that have driven the diversification of this genus: (1) linear progression of tusk form; (2) allopatric speciation through isolation in adjacent deep-sea canyons; (3) sympatric speciation through sexual selection on tusks; and (4) selection for species-recognition cues. Maximum-likelihood and Bayesian reconstructions confirmed the monophyly of the genus and revealed that what were considered ancestral and derived tusk forms have in fact arisen independently on several occasions, contrary to predictions from the linear-progression hypothesis. Further, none of the three well-supported species clades was confined to a single ocean basin, as might have been expected from the deep-sea canyon-isolation or sexual-selection hypotheses, and some species with similar tusks have overlapping distributions, contrary to predictions from the species-recognition hypothesis. However, the divergent tusk forms and sympatric distributions of three of the four sister-species pairs identified suggest that sexual selection on male tusks has likely played an important role in this unique radiation, although other forces are clearly also involved. To our knowledge, this is the first time that sexual selection has been explicitly implicated in the radiation of a mammalian group outside terrestrial ungulates.

Evolutionary radiations involving ornaments or displays have been documented in a wide range of organisms, including Hawaiian Drosophila flies (Kaneshiro, 1983), African Great Lake cichlids (Seehausen and van Alphen, 1999), and passerine birds (West-Eberhard, 1983). Radiations based on weapons of male-male competition are less common but also occur, though reasons why such characters might diverge in form are far less clear (Emlen et al., 2005). Examples of radiations of weaponry include the antlers and horns of ungulates (Geist, 1978; Lundrigan, 1996), the claws of amphipods and isopods (Shuster and Wade, 2003), the tusks of frogs (Shine, 1979), and the horns of dung and rhinoceros beetles (Eberhard, 1980; Emlen et al., 2005). In many species, secondary sexual characters can also fulfill dual functions (Berglund et al., 1996), as weapons in male-male competition (intra-sexual selection) and as cues for female choice (inter-sexual selection) or species recognition (see also below). Such traits often arise through male-male competition and serve as honest signals to other males regarding fighting ability or dominance. These traits may then be co-opted by females as indicators of male phenotypic quality when selecting a mate. Conversely, males may use information from traits that initially evolved through female choice (Berglund et al., 1996).

With 21 species described to date, the Ziphiidae (beaked whales) are the most speciose family in the order Cetacea after the Delphinidae (oceanic dolphins), yet they are among the least known of mammalian groups (Wilson, 1992). Beaked whales are deep-diving odontocetes (toothed whales) that live in the offshore waters of all the world's oceans except the highest latitudes of the Arctic. They are rarely seen at sea due to their elusive habits, long dive capacity, and, for some species, probable low abundance (Reeves et al., 2002). Most information has come from beachcast or stranded animals, and several species are known from only a handful of specimens.

The majority of species in this family comprise the genus Mesoplodon (n = 14); the other five recognized genera are monotypic or consist of antitropical species pairs. Mesoplodon beaked whales are relatively similar in overall appearance but differ dramatically from one another in their tooth morphology (Mead, 1989). These whales possess only a single pair of teeth, which vary by species in their shape, size, and position in the lower jaw. The tusk-like teeth develop only in adult males and protrude outside the mouth when the jaw is closed (Fig. 1). In females, the teeth do not develop and remain hidden in the gum tissues, so that the animals are effectively toothless. Examination of stomach contents indicates that these whales feed primarily on deep-water squid. The teeth do not appear to be used for feeding, and the small amount of evidence gathered to date suggests that there is little difference in diet between males and females (Sekiguchi et al., 1996). Instead, analysis of scarring patterns suggests that males use their tusks as weapons in intra-sexual combat, presumably to obtain reproductive access to females (Heyning, 1984). Several species of Mesoplodon are found in each ocean basin. On a broad scale, these species overlap in distribution (geographic sympatry) but may be ecologically parapatric due to differences in preferred prey (niche partitioning).

Figure 1

Adult male Gray's beaked whales, Mesoplodon grayi, stranded in New Zealand showing distinctive tusk-like teeth and linear scars likely caused by tooth rakes received in aggressive encounters with other males. Photocredits: G. Lento (upper), A. Glaser (lower).

Figure 1

Adult male Gray's beaked whales, Mesoplodon grayi, stranded in New Zealand showing distinctive tusk-like teeth and linear scars likely caused by tooth rakes received in aggressive encounters with other males. Photocredits: G. Lento (upper), A. Glaser (lower).

Four hypotheses have been put forward to explain patterns of Mesoplodon tusk morphology and/or the processes that have driven the diversification of this genus. After first clarifying the taxonomic status of many ziphiids (Moore, 1957, 1960, 1963, 1966), Moore (1968) proposed a linear progression of phenetic relationships among Mesoplodon species based primarily on the position and angle of inclination of the male tusks (Fig. 2). In other ziphiid genera, the tusks or largest pair of mandibular teeth are conical and occur at the apex of the lower jaw. Moore (1968) considered Mesoplodon to be the most derived of the extant genera, and argued that Mesoplodon species with apical, forward-inclined tusks were “primitive,” whereas those with backwards-inclined tusks set back from the apex of the mandible were more derived. This hypothesis provides a clear framework for the expected phylogenetic relationships among the species but makes no prediction regarding patterns of geographic distribution and provides no mechanism to explain the process by which speciation occurred.

Figure 2

Adult male Mesoplodon beaked whale tusk morphology (lateral views of lower jaws), arranged in a linear progression from ancestral to derived, as perceived by Moore (1968) based on a combination of tusk position, size, and angle of inclination (hypothesis H1-A). Other beaked whale genera have forwardly inclined tusks at the apex of the lower jaw. Although the tusks of mature males are, in general, accurately portrayed in these images, the potential for and degree of intraspecific variation that may occur in these characters has not been shown. Note that tusk coding options (Fig. 3, Table 3) were based on quantitative data from the literature and not reliant on the images shown here. Jaw images are labeled with three-or four-letter species codes, scientific names, and common names. Asterisks highlight species discovered since 1968. For M. peruvianus, the circle highlights the position of the small tooth. M. traversii, a species with teeth similar to those of M. layardii (van Helden et al. 2002), but not included in this genetic study, is not shown.

Figure 2

Adult male Mesoplodon beaked whale tusk morphology (lateral views of lower jaws), arranged in a linear progression from ancestral to derived, as perceived by Moore (1968) based on a combination of tusk position, size, and angle of inclination (hypothesis H1-A). Other beaked whale genera have forwardly inclined tusks at the apex of the lower jaw. Although the tusks of mature males are, in general, accurately portrayed in these images, the potential for and degree of intraspecific variation that may occur in these characters has not been shown. Note that tusk coding options (Fig. 3, Table 3) were based on quantitative data from the literature and not reliant on the images shown here. Jaw images are labeled with three-or four-letter species codes, scientific names, and common names. Asterisks highlight species discovered since 1968. For M. peruvianus, the circle highlights the position of the small tooth. M. traversii, a species with teeth similar to those of M. layardii (van Helden et al. 2002), but not included in this genetic study, is not shown.

The second hypothesis seeks to explain the overall morphological similarity of beaked whales and suggests that speciation was driven by population isolation in deep-sea canyons (Carwardine, 1995). Beaked whales rarely venture into shallow waters and most sightings occur beyond the edge of the continental shelf, often associated with submarine canyons and seamounts (Barlow et al., 2006; MacLeod et al., 2006). Therefore, populations in different canyon systems may be largely isolated from one another, yet would evolve in parallel to adapt to similar conditions. Under this form of allopatric speciation, species in adjacent deep-water regions within an ocean basin would be expected to be more closely related to one another than to species in other ocean basins. However, the canyon-isolation hypothesis makes no predictions about tusk morphology.

The third hypothesis proposes that sexual selection may have been an important driver of speciation and diversification in this genus (Dalebout, 2002). This idea is supported by the strong sexual dimorphism in Mesoplodon tooth morphology and its presumed role in intra-sexual competition (Heyning, 1984). Under the sexual-selection hypothesis, directional or disruptive selection on tooth morphology could lead to sympatric speciation. This predicts that groups of related species that differ in tusk morphology would co-occur in the same ocean basin and, specifically, that sister-species with similar distributions are likely to have highly divergent tooth forms.

The fourth hypothesis suggests that the variation observed in the position and shape of the tusks could function as species recognition cues and therefore as a precopulatory isolating mechanism (MacLeod, 2000). Although this hypothesis does not address the process by which speciation initially occurred, it predicts that divergent tooth morphologies would arise through directional selection and character displacement in areas of secondary contact between species. This process would therefore also result in divergent tooth morphology among species with overlapping distributions.

Testing the predictions of these hypotheses requires a robust phylogeny of Mesoplodon species. However, since Moore's (1968) pioneering morphological work, there has been little further investigation of beaked whale evolutionary relationships. Several large-scale studies of cetacean phylogeny have included a handful of ziphiids but revealed little about relationships within this family (Arnason and Gullberg, 1996; Messenger and McGuire, 1998; Cassens et al., 2000; May-Collado and Agnarsson, 2006). All 21 known ziphiid species were recently confirmed to be genetically distinct based on mitochondrial (mt) DNA control region and cytochrome b sequences (Dalebout et al., 2004). To determine whether patterns of species distinctiveness were reflected in the nuclear genome as well, Dalebout et al. (2004) also compiled a suite of actin intron sequences. Although higher-level relationships were not well-resolved by the rapidly evolving mtDNA markers due to phylogenetic noise and saturation (see also Dalebout et al., 2002, 2003), preliminary analysis of the intron sequences indicated that a multiple nuclear gene approach should produce a robust and highly resolved phylogeny for this remarkable genus.

Here we use partial sequences from seven nuclear introns (3348 base pairs) to construct such a phylogeny for the majority of Mesoplodon beaked whales. The resulting nuclear phylogeny was used as a framework within which to test predictions from the four hypotheses of speciation and tusk evolution (Fig. 3): the linear-progression hypothesis (H1) predicts that tusk morphology will follow a trend from ancestral to more derived forms; both the deep-sea canyon-isolation (H2) and the sexual-selection hypotheses (H3) predict that clades of related species will occur in the same ocean basin; and both H3 and the species-recognition hypothesis (H4) predict that species with overlapping or adjacent distributions will have divergent tusk forms. However, H3 predicts that tusk divergence will be most pronounced in sister-species pairs (sympatric speciation), whereas for H4, the species involved need not necessarily be so closely related (i.e., secondary contact).

Figure 3

Predictions from the four hypotheses: H1, linear progression of tusk morphology; H2, deep-sea canyon isolation; H3, sexual selection; and H4, species recognition. See text for discussion. For H1, the column highlighted in gray (A) represents the linear progression in tusk form proposed by Moore (1968) based on a combination of tusk position, size, and angle of inclination in the jaw as shown in Figure 2. See Table 3 for details of this and other (B through J) tusk coding options. The H1 table and terminal nodes on the H2 tree are labeled with species codes as per Figure 2. For H2 and H3, shaded boxes and circles indicate species' distributions: North Atlantic, white; North Pacific, gray; Southern Hemisphere, hatched. For species found in more than one ocean basin, the main center of distribution was used, such that all species, except M. densirostris (found in all oceans), could be assigned to a specific ocean-basin clade (i.e., equivalent to the H2+H3 “relaxed” option in Table 5).

Figure 3

Predictions from the four hypotheses: H1, linear progression of tusk morphology; H2, deep-sea canyon isolation; H3, sexual selection; and H4, species recognition. See text for discussion. For H1, the column highlighted in gray (A) represents the linear progression in tusk form proposed by Moore (1968) based on a combination of tusk position, size, and angle of inclination in the jaw as shown in Figure 2. See Table 3 for details of this and other (B through J) tusk coding options. The H1 table and terminal nodes on the H2 tree are labeled with species codes as per Figure 2. For H2 and H3, shaded boxes and circles indicate species' distributions: North Atlantic, white; North Pacific, gray; Southern Hemisphere, hatched. For species found in more than one ocean basin, the main center of distribution was used, such that all species, except M. densirostris (found in all oceans), could be assigned to a specific ocean-basin clade (i.e., equivalent to the H2+H3 “relaxed” option in Table 5).

Materials and Methods

DNA Extraction, Amplification, and Sequencing

We sampled 13 of the 14 known Mesoplodon species and two confamilial outgroups (Ziphius cavirostris and Tasmacetus shepherdi). Soft tissue samples were obtained from dead, stranded animals or victims of incidental fisheries takes (bycatch) and preserved in 70% ethanol or 20% salt-saturated dimethyl sulphoxide (DMSO) and stored at 4°C or −20°C prior to genetic analysis. The missing species, M. traversii, is currently known only from fragmentary skeletal material (van Helden et al., 2002), which is unsuitable for amplification of nuclear introns. DNA was extracted using standard phenol:chloroform methods (Sambrook et al., 1989), as modified for small samples by Baker et al. (1994). The polymerase chain reaction (PCR) was used to amplify partial introns from seven nuclear genes: biglycan (BGN), 706 base pairs (bp); catalase (CAT), 559 bp; rhodopsin (RHO), 166 bp; cytotoxic T-lymphocyte-associated serine esterase 3 (CTLA3), 305 bp; cholinergic receptor–nicotinic, alpha polypeptide 1 (CHRNA1), 366 bp; muscle actin (ACT), 925 bp; and major histocompatibility complex class II (DQA), 456 bp. Each of the seven genes are found on a different chromosome in humans (Lyons et al., 1997) and are assumed to be unlinked and independent in cetaceans. As the vast majority of genes are unlinked (e.g., Turelli, 1984), we considered this a reasonable assumption. See Table 1 for primer and PCR information. PCR products were sequenced on an ABI 377, modified ABI 373, or ABI 3700 automated sequencer (Applied Biosystems, Inc.) using BigDye Dye Terminator Chemistry. Fragments were sequenced at least twice in both directions for confirmation in the majority of cases. Sequences were aligned and edited manually using the program SEQUENCHER ver. 4.0 (Gene Codes Corporation, Inc.). Alignment of intron sequences was generally straightforward as relatively few insertions and deletions appear to have occurred since their divergence from common ancestral sequences. GenBank accession numbers for all sequences are listed in Table 2.

Table 1

Primers used for amplification of nuclear introns in beaked whales. All PCR reactions used a final concentration of 2.5 mM Mg2 +. Tm, annealing temperature in °C.

Primer name Sequence Primer source Tm 
BGN-F 5′-CTCCAAGAACCACCTGGTG-3′ Lyons et al. (1997) 62 
BGN-R 5′-TTCAAAGCCACTGTTCTCCAG-3′ Lyons et al. (1997)  
CAT-F 5′-AAAGACTGACCAGGGCATCA-3′ Lyons et al. (1997) 55 
CAT-R 5′-AGGGTAGTCCTTGTGAGGCC-3′ Lyons et al. (1997)  
RHO-F 5′-AGGGGAGGTCACTTTATAAGGG-3′ Lyons et al. (1997) 55 
RHO-R 5′-CCAGCATGGAGAACTGCC-3′ Lyons et al. (1997)  
CTLA3-F 5′-AAGAATTTCCCTATCCATGCTATG-3′ Lyons et al. (1997) 50 
CTLA3-R 5′-GGTTCCTGGTTTCACATCATC-3′ Lyons et al. (1997)  
CHRNA1-F 5′-GACCATGAAGTCAGACCAGGAG-3′ Lyons et al. (1997) 55 
CHRNA1-R 5′-GGAGTATGTGGTCCATCACCAT-3′ Lyons et al. (1997)  
ACT3-Fa 5′-GGTTATCTGATGTATTCC-3′ Palumbi and Baker (1994) 50 
ACT1385-Ra 5′-CTTGTGAACTGATTACAGTCC-3′ Palumbi and Baker (1994)  
ACT5-Fa 5′-CCACTACTTTAGGCAG-3′ Palumbi and Baker, unpublished  
M13-ACT5-Ra 5′-TGTAAAACGACGGCCAGTCTGCCTAAAGTAGTGG-3′ Palumbi and Baker, unpublished  
DQA1-F 5′-CCGGATCCCAGTACACCCATGAATTTGATGG-3′ Auffray et al. (1987) 54 
DQA2-R 5′-CCGGATCCCCAGTGCTCCACCTTGCAGTC-3′ Auffray et al. (1987)  
Primer name Sequence Primer source Tm 
BGN-F 5′-CTCCAAGAACCACCTGGTG-3′ Lyons et al. (1997) 62 
BGN-R 5′-TTCAAAGCCACTGTTCTCCAG-3′ Lyons et al. (1997)  
CAT-F 5′-AAAGACTGACCAGGGCATCA-3′ Lyons et al. (1997) 55 
CAT-R 5′-AGGGTAGTCCTTGTGAGGCC-3′ Lyons et al. (1997)  
RHO-F 5′-AGGGGAGGTCACTTTATAAGGG-3′ Lyons et al. (1997) 55 
RHO-R 5′-CCAGCATGGAGAACTGCC-3′ Lyons et al. (1997)  
CTLA3-F 5′-AAGAATTTCCCTATCCATGCTATG-3′ Lyons et al. (1997) 50 
CTLA3-R 5′-GGTTCCTGGTTTCACATCATC-3′ Lyons et al. (1997)  
CHRNA1-F 5′-GACCATGAAGTCAGACCAGGAG-3′ Lyons et al. (1997) 55 
CHRNA1-R 5′-GGAGTATGTGGTCCATCACCAT-3′ Lyons et al. (1997)  
ACT3-Fa 5′-GGTTATCTGATGTATTCC-3′ Palumbi and Baker (1994) 50 
ACT1385-Ra 5′-CTTGTGAACTGATTACAGTCC-3′ Palumbi and Baker (1994)  
ACT5-Fa 5′-CCACTACTTTAGGCAG-3′ Palumbi and Baker, unpublished  
M13-ACT5-Ra 5′-TGTAAAACGACGGCCAGTCTGCCTAAAGTAGTGG-3′ Palumbi and Baker, unpublished  
DQA1-F 5′-CCGGATCCCAGTACACCCATGAATTTGATGG-3′ Auffray et al. (1987) 54 
DQA2-R 5′-CCGGATCCCCAGTGCTCCACCTTGCAGTC-3′ Auffray et al. (1987)  
a

See Dalebout et al. (2004) for ACT primer map. ACT5-F and M13-ACT5-R are internal primers within the large fragment amplified by ACT3-F and ACT1385-R and were used for sequencing only.

Table 2

The seven nuclear introns analysed for this study. The number of base pairs (bp) analyzed for each gene segment, GenBank accession numbers, the number of variable and phylogenetically informative (PI) sites among all Mesoplodon species and among Mesoplodon and outgroup species (n = 2), the mean pairwise genetic distances (Tamura 3-parameter model, gamma rates [alpha = 1.0]) with standard errors (SE) among Mesoplodon species, and transition/transversion ratios (Ti/Tv) are shown.

   No. variable sites  Mean pairwise Mean 
 Number bp GenBank accession Mesoplodon/all No. PI sites distance (SE) Ti/Tv(SE) 
Gene analyzed numbers taxa Mesoplodon/all taxa Mesoplodon Mesoplodon 
BGN 671 EU447744–EU447758 27/36 6/11 0.009 (0.002) 2.4 (0.77) 
CAT 516 EU447759–EU447773 10/12 3/4 0.004 (0.001) 1.1 (0.53) 
RHO 166 EU476106–EU476120 4/8 0/3 0.009 (0.002) 0.3 (0.15) 
CTLA3 264 EU476121–EU476135 10/20 4/8 0.008 (0.003) 1.0 (0.63) 
CHRNA1 350 EU476136–EU476150 8/16 0/1 0.003 (0.001) 0.2 (0.40) 
ACT 925 EU476151–EU476165 37/44 5/9 0.008 (0.002) 3.4 (0.91) 
DQA 456 EU476166–EU476180 18/25 9/9 0.008 (0.002) 2.7 (1.02) 
COMBO 3348  114/161 27/45 0.007 (0.001) 2.0 (0.60) 
   No. variable sites  Mean pairwise Mean 
 Number bp GenBank accession Mesoplodon/all No. PI sites distance (SE) Ti/Tv(SE) 
Gene analyzed numbers taxa Mesoplodon/all taxa Mesoplodon Mesoplodon 
BGN 671 EU447744–EU447758 27/36 6/11 0.009 (0.002) 2.4 (0.77) 
CAT 516 EU447759–EU447773 10/12 3/4 0.004 (0.001) 1.1 (0.53) 
RHO 166 EU476106–EU476120 4/8 0/3 0.009 (0.002) 0.3 (0.15) 
CTLA3 264 EU476121–EU476135 10/20 4/8 0.008 (0.003) 1.0 (0.63) 
CHRNA1 350 EU476136–EU476150 8/16 0/1 0.003 (0.001) 0.2 (0.40) 
ACT 925 EU476151–EU476165 37/44 5/9 0.008 (0.002) 3.4 (0.91) 
DQA 456 EU476166–EU476180 18/25 9/9 0.008 (0.002) 2.7 (1.02) 
COMBO 3348  114/161 27/45 0.007 (0.001) 2.0 (0.60) 

Phylogenetic Analyses

Files for each of the seven introns were concatenated to form a combined data set (COMBO, 3348 bp) using MacClade ver. 4.07 (Maddison and Maddison, 2005). We attempted to use the same set of individuals to generate all sequences, but amplification problems with poor-quality DNA from decomposing stranded specimens required us to create chimeric sequences in some cases (see Appendix 1). Data on variable sites, shared-derived (phylogenetically informative [PI]) sites, estimates of transition/transversion ratio, and the mean divergence among ingroup taxa for each gene were obtained using MEGA3 (Kumar et al., 2004).

Phylogenetic analysis of individual introns and the COMBO data set was performed using maximum parsimony (MP), minimum evolution (ME; neighbor joining), and maximum-likelihood (ML) methods as implemented in PAUP* 4.0b10 (Swofford 2003). To test for conflicting signal among gene segments, an initial evaluation of the individual introns was performed with a bootstrap support/conflict criterion of 90% (de Queiroz, 1993; Teeling et al., 2005). MP-based pairwise partition homogeneity tests (Farris et al., 1995) with 1000 replicates were also performed. MP analyses of the COMBO data set were conducted using random addition of taxa, tree-bisection reconnection branch swapping, without considering insertion-deletions (indels) as fifth characters, and a limit of 1,000,000 rearrangements. For ME and ML analyses, the Akaike information criterion (AIC) (Akaike, 1973; Posada and Buckley, 2004) was used to select a K81uf+I+G model with estimated proportion of invariable sites (Pinvar, I) = 0.7368 and gamma (G)-shape parameter for distribution of rates (alpha) = 0.9307, using ModelTest Online ver. 3.8 (Posada and Crandall, 1998; Posada, 2006). For comparison, additional analyses were run using a Bayesian information criterion (BIC)-selected model (HKY+I [Pinvar, I = 0.8449]). For ML, starting trees were obtained via neighbor joining. The reliability of nodes was assessed using 500 full heuristic (MP, ML) or 1000 neighbor-joining (ME), non-parametric bootstrap replicates. Clades with bootstrap values ≥70% were considered robust.

Bayesian (BAY) analyses of the COMBO data set were performed using MrBayes ver. 3.1.2 (Huelsenbeck and Ronquist, 2001; Ronquist and Huelsenbeck, 2003). Analyses were conducted using uniform priors, random starting trees, no phylogenetic constraints, and four simultaneous Markov chains run for 10,000,000 generations, with trees sampled every 100 generations and the first 250,000 generations discarded as burn-in. The ML model for BAY analyses used two substitution types, with rate variation across sites estimated using the “invgamma” option, such that a proportion of the sites is invariable (Pinvar), whereas the rate for the remaining sites (alpha) is drawn from a gamma distribution. Each BAY run was replicated to ensure a convergence of results. Assessment of convergence was based on several diagnostics: the average standard deviation of split frequencies (less than 0.002 at the end of a run), a plot of generation versus log-likelihood scores (at convergence, different independent runs are expected to sample similar likelihood values resulting in a “white noise” plot), and all PSRF (potential scale reduction factor) scores are equal to 1.000 (Ronquist et al., 2005). An extra-long run (25,000,000 generations) was also conducted to ensure that a robust convergence had been reached. Clades with posterior probabilities of ≥0.95 were considered robust (Rannala and Yang, 1996).

To test whether the strength of inferences of relationships among ingroup taxa was influenced by choice of outgroups, COMBO analyses were repeated using only one of the two available outgroups and the resulting topologies compared. For the same purpose, an additional set of analyses (MP, ME, NJ, and BAY) was conducted using a reduced combined data set, including six of the seven introns (3084 bp; missing CTLA3) and a third confamilial outgroup (Hyperoodon planifrons).

Testing Alternative Hypotheses

To test predictions from the four hypotheses, tusk forms were sorted into categories following the linear progression proposed by Moore (1968) based on a combination of tusk position, size, and angle of inclination (Fig. 2 and Fig. 3, H1-A), where 1 = apical; 2 = larger tusks set some way back from apex; 3 = massive triangular tusks in middle of jaw; 4 = tusks in middle of jaw with strong backwards inclination; and 5 = tusks raised up on prominent arch with a forwards inclination. To test the sensitivity of these categories, nine other coding schemes (H1-B to H1-J) based on various tusk characters assumed to be more independent (e.g., position, size, shape, angle of inclination, on arch or not) were also assessed (Table 3, Fig. 3). Predictions from hypotheses H1 to H3 were compared statistically under ML to the best phylogenetic tree using approximately unbiased tests (AU; Shimodaira, 2002) and Shimodaira-Hasegawa tests (SH; Shimodaira and Hasegawa, 1999), as implemented in the program CONSEL ver. 0.1i (Shimodaira and Hasegawa, 2001). Both these tests are appropriate for comparing both a priori and a posteriori phylogenetic hypotheses, unlike the Kishino and Hasegawa (1989) test. The SH test may, however, be overly conservative because the number of trees included in the confidence set increases as the number of trees being considered increases (Strimmer and Rambaut, 2002). The AU test uses a multi-scale bootstrap technique to remove this bias and has been recommended for general tree selection problems (Shimodaira, 2002). Predictions from hypotheses H3 and H4 were assessed using ML reconstructions of ancestral states (distribution, tusk form) as implemented in the program MESQUITE (Maddison and Maddison, 2006). We also considered using the program BayesTraits Multistate (Pagel and Meade, 2006), but unfortunately this application is unable to deal with polytomies. A single outgroup (Ziphius cavirostris) was used for these and subsequent analyses (see below).

Table 3

Details of Mesoplodon tusk coding options.

H1-A Linear progression 1 = apical 
   in tusk form based 2 = larger tusks set some way back from apex 
   on combination of 3 = massive triangular tusks in middle of jaw 
   tusk position, size, 4 = tusks in middle of jaw with strong backwards inclination 
   and angle of 5 = tusks raised up on prominent arch and angled forwards 
   inclination in jaw  
   (Moore 1968 
H1-B Tusk position—as 1 = apical 
   qualitative measurea 2 = intermediate distance 
  3 = furthest back 
H1-C Tusk position—in 1 = apical 
   relation to 2 = between apex and posterior end of symphysis 
   mandibular 3 = overlapping with posterior end of symphysis 
   symphysisb 4 = posterior to mandibular symphysis 
H1-D Tusk position—as 1 = apical 
   quantitative 2 = approx. 100 mm back 
   measurec 3 = approx. 200 mm back 
  4 = approx. 300 mm back 
H1-E Tusk size 1 = small 
  2 = medium 
  3 = massive 
  4 = straplike 
H1-F Tusk lengthb, d 1 = less than 100 mm 
  2 = 100–200 mm 
  3 = greater than 200 mm 
H1-G Tusk widthb, d 1 = less than 50 mm 
  2 = greater than 60 mm 
  3 = greater than 70 mm 
  4 = greater than 100 mm 
H1-H Tusk shape 1 = approx. conical 
  2 = flattened, triangular, small 
  3 = flattened, triangular, large 
  4 = straplike 
H1-I Tusk inclination 1 = anterior 
  2 = vertical 
  3 = mild posterior 
  4 = strong posterior 
H1-J Tusk raised up on prominent arch 1 = yes 
  2 = no 
H1-A Linear progression 1 = apical 
   in tusk form based 2 = larger tusks set some way back from apex 
   on combination of 3 = massive triangular tusks in middle of jaw 
   tusk position, size, 4 = tusks in middle of jaw with strong backwards inclination 
   and angle of 5 = tusks raised up on prominent arch and angled forwards 
   inclination in jaw  
   (Moore 1968 
H1-B Tusk position—as 1 = apical 
   qualitative measurea 2 = intermediate distance 
  3 = furthest back 
H1-C Tusk position—in 1 = apical 
   relation to 2 = between apex and posterior end of symphysis 
   mandibular 3 = overlapping with posterior end of symphysis 
   symphysisb 4 = posterior to mandibular symphysis 
H1-D Tusk position—as 1 = apical 
   quantitative 2 = approx. 100 mm back 
   measurec 3 = approx. 200 mm back 
  4 = approx. 300 mm back 
H1-E Tusk size 1 = small 
  2 = medium 
  3 = massive 
  4 = straplike 
H1-F Tusk lengthb, d 1 = less than 100 mm 
  2 = 100–200 mm 
  3 = greater than 200 mm 
H1-G Tusk widthb, d 1 = less than 50 mm 
  2 = greater than 60 mm 
  3 = greater than 70 mm 
  4 = greater than 100 mm 
H1-H Tusk shape 1 = approx. conical 
  2 = flattened, triangular, small 
  3 = flattened, triangular, large 
  4 = straplike 
H1-I Tusk inclination 1 = anterior 
  2 = vertical 
  3 = mild posterior 
  4 = strong posterior 
H1-J Tusk raised up on prominent arch 1 = yes 
  2 = no 
a

Data from Moore (1968).

b

Data from Mead (1989).

c

Data from Carwardine (1995).

Estimation of Divergence Dates

A likelihood-ratio test was performed using PAUP* to assess if substitution rates could be modeled as clocklike. Based on these results, PAML ver. 3.14 (Yang, 1997) and ESTBRANCHES (Thorne et al., 1998; Kishino et al., 2001) were used to estimate branch lengths for the COMBO data set ML topology, using a F84+G model with five rate categories, following the recommendations of Rutschmann (2005). The likelihood scores obtained were compared to check convergence. Divergence times were then estimated using MultiDivTime (Thorne et al., 1998), a Bayesian approach that does not assume a strict clock and can incorporate user-provided constraints.

There are only a limited number of fossil calibration points for beaked whales. The family arose in the early Miocene (∼ 24 million years ago [Ma]), reached the height of its morphological diversity in the mid-Miocene, and remained well represented in late Miocene (7 to 10 Ma; Barnes et al., 1985; Mead, 1989). Unfortunately, the majority of ziphiid fossils lack the diagnostic characters required for attribution to extant lineages, and the few that do possess such features have generally been obtained from geological strata that cannot be dated (e.g., Owen, 1870; Kellogg, 1928a, 1928b). As a result, the origin of the modern genera remains in doubt (Mead, 1989; Fordyce, 2002). However, an estimate of 24 ± 5 Ma for the divergence of Ziphius and Mesoplodon was provided by a molecular investigation of phylogenetic relationships among ancient cetacean lineages (Cassens et al., 2000—as inferred from Fig. 2; a multi-gene phylogeny calibrated using fossil dates for the origin of the modern family Delphinidae, 11 to 13 Ma). Further, a number of specimens attributable to the Ziphius lineage have recently been described from late Miocene deposits (approximately 10 Ma; Lambert, 2005).

We therefore explored divergence dates obtained from several combinations of parameters that constrained the origin of the genus Mesoplodon to 30 Ma (option A), 20 Ma (option B), and 10 Ma (option C). Two versions of each option were run: the first using only an upper bound (maximum age; 30, 20, and 10 Ma, respectively), and the second using both an upper and lower bound (± 5 Ma around maximum ages used for the first). Following recommendations in Thorne's readme file (http://statgen.ncsu.edu/thorne/multidivtime.html), and using 10 Ma as one time unit, values used for the a priori number of expected time units between tip and root (ingroup depth; rttm), the mean of prior distribution for rate at root node (rtrate), and the prior for the Brownian motion parameter (nu) were as follows: (A) 3.0, 0.001, 0.3; (B) 2.0, 0.001, 0.5; and, (C) 1.0, 0.002, 1.0. Bigtime was set to 50 Ma. After a burn-in of 100,000 generations, 1,000,000 generations of Markov chains were sampled every 100 generations. Posteriors for each option were computed twice (using different random starting points) to check convergence of results.

Results

Variation in Phylogenetic Signal among Introns

The seven introns differed in their contribution to the phylogenetic signal, as reflected in the number of shared-derived (phylogenetically informative, PI) sites (Table 2), ranging from 0 sites (RHO and CHRNA1) to 9 sites (DQA), with a total of 27 PI sites for the COMBO data set. The amount of phylogenetic signal had some correlation with segment length (bp), with longer introns generally contributing more PI sites. Pairwise genetic distances among Mesoplodon species also varied by intron segment, suggesting some differences in the rate of accumulation of mutations among loci. There was little resolution of relationships in the individual intron analyses (see Appendix 2). Of the few strongly supported nodes, only one was not observed in subsequent COMBO reconstructions (BGN, 85% bootstrap support). Partition homogeneity tests revealed low-level conflict (P = 0.03–0.05) in phylogenetic signal between BGN and three other introns (CTLA3, ACT, and DQA), as well as between CTLA3 and ACT and between ACT and DQA. Serial removal and replacement of single species from the analyses did not affect these results, suggesting that the conflict was locus-rather than taxon-specific. However, none of the pairwise tests were significant after Bonferroni corrections for multiple comparisons. This overall lack of conflict among individual introns also provides some support for our assumption that these genes are unlinked and independent in cetaceans.

Phylogenetic Reconstruction with Combined Intron Data Set

In contrast to the individual intron analyses, the COMBO data set yielded a robust, highly resolved phylogeny. Three main clades were strongly supported by bootstrap scores (BS) and/or Bayesian posterior probabilities (BPP; Fig. 4). The same clades with similar levels of high support were obtained when BGN was excluded and with different combinations of outgroups. Further support for the phylogeny was provided by lineage-specific, shared-derived nucleotide substitutions and deletions (synapomorphies; Table 4).

Figure 4

Phylogenetic relationships among Mesoplodon beaked whale species reconstructed using maximum likelihood and an AIC-selected model (−ln L = 6055.03782). The same robust topology was obtained from analyses using a BIC-selected model. Bootstrap scores ≥50% are shown above or alongside internal nodes (MP/ME/ML), and Bayesian posterior probabilities ≥0.80 are shown below. Node E was collapsed into a polytomy by MP, ME, and ML bootstrap analyses (i.e., bootstrap scores < 50%) but was strongly supported by BAY analysis (0.99). See Table 4 for distribution of shared-derived sites supporting these nodes. <, less than 50% bootstrap score.

Figure 4

Phylogenetic relationships among Mesoplodon beaked whale species reconstructed using maximum likelihood and an AIC-selected model (−ln L = 6055.03782). The same robust topology was obtained from analyses using a BIC-selected model. Bootstrap scores ≥50% are shown above or alongside internal nodes (MP/ME/ML), and Bayesian posterior probabilities ≥0.80 are shown below. Node E was collapsed into a polytomy by MP, ME, and ML bootstrap analyses (i.e., bootstrap scores < 50%) but was strongly supported by BAY analysis (0.99). See Table 4 for distribution of shared-derived sites supporting these nodes. <, less than 50% bootstrap score.

Table 4

Summary of the distribution of sites supporting internal nodes in the phylogeny as labeled in Figure 4. All characters are single-nucleotide substitutions unless indicated otherwise. Numbers in parentheses indicate how many of these represent unique, lineage-specific, shared-derived characters (synapomorphies)—i.e., changes that occur only once in the tree.

Node BGN CAT RHO CTLA3 CHRNA1 ACT DQA Total 
4 (4) 1 (1) 3 (3) 3 (2) 1 (1) 1 (1)  13 (12) 
     3 (3) 1 (0) 4 (3) 
     2 (1)  2 (1) 
1 (0)       1 (0) 
   2 (0)   1 (0) 3 (0) 
   1a (1)  2 (2)  3 (3) 
       0 (0) 
     1 (1)  1 (1) 
 1 (1)  1 (1)   1 (1) 3 (3) 
2 (0) 1 (0)     1 (1) 4 (1) 
1 (1)      3 (1) 4 (2) 
 8 (5) 3 (2) 3 (3) 7 (4) 1 (1) 9 (8) 7 (3) 38 (26) 
Node BGN CAT RHO CTLA3 CHRNA1 ACT DQA Total 
4 (4) 1 (1) 3 (3) 3 (2) 1 (1) 1 (1)  13 (12) 
     3 (3) 1 (0) 4 (3) 
     2 (1)  2 (1) 
1 (0)       1 (0) 
   2 (0)   1 (0) 3 (0) 
   1a (1)  2 (2)  3 (3) 
       0 (0) 
     1 (1)  1 (1) 
 1 (1)  1 (1)   1 (1) 3 (3) 
2 (0) 1 (0)     1 (1) 4 (1) 
1 (1)      3 (1) 4 (2) 
 8 (5) 3 (2) 3 (3) 7 (4) 1 (1) 9 (8) 7 (3) 38 (26) 
a

Two-base pair deletion.

The radiation of genus Mesoplodon appears to have begun with the divergence of the lineage leading to Sowerby's beaked whale, M. bidens. This was followed by several rapid divergence events resulting in the straptoothed whale clade (M. layardii, M. bowdoini, and M. carlhubbsi; BS MP 61/ME 73/ML 68, BPP 0.99), the Perrin's beaked whale clade (M. hectori, M. peruvianus, M. perrini, M. grayi, M. stejnegeri, and M. densirostris; BS 60/59/60, BPP 0.99), and the True's beaked whale clade (M. ginkgodens, M. mirus, and M. europaeus; BS 90/93/87, BPP 1.00). Within the straptoothed whale clade, M. carlhubbsi and M. bowdoini were identified as likely sister species, though this node received somewhat lower support (BS < 50/69/ < 50, BPP 0.83). Within the Perrin's beaked whale clade, M. stejnegeri and M. densirostris and M. peruvianus and M. perrini were identified as sister species (BS 74/79/69, BPP 0.97, and BS 96/95/93, BPP 1.00), and within the True's beaked whale clade, M. mirus and M. europaeus were identified as sister species (BS 99/100/97, BPP 1.00).

Alternative Hypotheses of Evolution

AU and SH tests revealed that alternative hypotheses of relationships among Mesoplodon species predicted from H1 to H3 were significantly less likely than the best ML tree (P < 0.001; Table 5). Instead of following a linear progression, tusk form appears to be highly plastic, and forms considered to be ancestral or derived under the scheme proposed by Moore (1968) appear to have evolved independently on a number of occasions. Similarity in tusk morphology was generally not a good indicator of relatedness, and most sister species possessed divergent tusk forms (e.g., M. perrini and M. peruvianus). Further, none of the three main clades was confined to a single ocean basin, as might be expected from H2 and H3 (Fig. 5).

Figure 5

Maximum likelihood topology with estimated dates of divergence (asterisk highlights constraint node—origin of genus Mesoplodon between 25 and 15 million years ago). Terminal nodes are labeled with species codes as per Figure 2. Tusk morphology is depicted by jaw images adjacent to branch termini. See text for discussion of tusk category codes (T1 to T5). Shaded boxes indicate species' distributions: North Atlantic, white; North Pacific, gray; Southern Hemisphere, hatched. Some species occur in multiple ocean basins. Ma, million years ago; Ple, Pleistocene.

Figure 5

Maximum likelihood topology with estimated dates of divergence (asterisk highlights constraint node—origin of genus Mesoplodon between 25 and 15 million years ago). Terminal nodes are labeled with species codes as per Figure 2. Tusk morphology is depicted by jaw images adjacent to branch termini. See text for discussion of tusk category codes (T1 to T5). Shaded boxes indicate species' distributions: North Atlantic, white; North Pacific, gray; Southern Hemisphere, hatched. Some species occur in multiple ocean basins. Ma, million years ago; Ple, Pleistocene.

Table 5

Approximately unbiased (AU) and Shimodaira-Hasegawa (SH) test scores comparing the best maximum likelihood (ML) tree with alternative hypotheses. See Table 3 and Figure 3 for details of H1 tusk coding.

 Difference in   
 
 
  
 −ln L ln L AU P value SH P value 
ML best tree 6063.049 (best) 1.000 0.999 
H1-A 6196.247 133.198 < 0.0001 0.001 
H1-B 6209.074 146.025 < 0.0001 0.001 
H1-C 6162.888 99.839 < 0.0001 0.001 
H1-D 6208.283 145.234 < 0.0001 0.001 
H1-E 6204.027 140.978 < 0.0001 < 0.0001 
H1-F 6211.812 148.763 < 0.0001 0.001 
H1-G 6215.172 152.123 < 0.0001 0.001 
H1-H 6209.628 146.579 < 0.0001 0.001 
H1-I 6205.718 142.669 < 0.0001 < 0.0001 
H1-J 6211.714 148.665 < 0.0001 0.001 
H2+H3 stricta 6212.809 149.760 < 0.0001 < 0.0001 
H2+H3 relaxedb 6195.401 132.352 < 0.0001 < 0.0001 
 Difference in   
 
 
  
 −ln L ln L AU P value SH P value 
ML best tree 6063.049 (best) 1.000 0.999 
H1-A 6196.247 133.198 < 0.0001 0.001 
H1-B 6209.074 146.025 < 0.0001 0.001 
H1-C 6162.888 99.839 < 0.0001 0.001 
H1-D 6208.283 145.234 < 0.0001 0.001 
H1-E 6204.027 140.978 < 0.0001 < 0.0001 
H1-F 6211.812 148.763 < 0.0001 0.001 
H1-G 6215.172 152.123 < 0.0001 0.001 
H1-H 6209.628 146.579 < 0.0001 0.001 
H1-I 6205.718 142.669 < 0.0001 < 0.0001 
H1-J 6211.714 148.665 < 0.0001 0.001 
H2+H3 stricta 6212.809 149.760 < 0.0001 < 0.0001 
H2+H3 relaxedb 6195.401 132.352 < 0.0001 < 0.0001 
a

Only species found in a single ocean basin were assigned to ocean-basin clades. See Figure 5 for information on species distributions.

b

For species found in more than one ocean basin, the main center of distribution was used to designate ocean basin of origin, such that all species, except M. densirostris, were assigned to ocean-basin clades.

Reconstruction of ancestral states confirmed the high plasticity of tusk morphology with all five tusk forms under H1-A estimated to be approximately equally likely at all internal nodes (Fig. 6a, right). Similar plasticity was observed in assessments of more independent features of tusk morphology (Fig. 6b, H1-B to H1-J). In only one case was a specific tusk character state found to be diagnostic for a particular clade; only members of the straptoothed whale clade, together with the ancestral lineage M. bidens, have tusks that overlap with the posterior end of the mandibular symphysis (H1-C, state 3). Better resolution was obtained from the ancestral area analysis (Fig. 6a, left): the straptoothed whale clade most likely arose in the Southern Hemisphere before dispersing into the North Pacific; the Perrin's beaked whale clade most likely arose in the Southern Hemisphere before dispersing into the North Pacific and beyond; and the True's beaked whale clade most likely arose in the North Pacific before dispersing into the North Atlantic. It should be recognized, however, that the ML model implemented in MESQUITE is quite simple and, as such, caution should be exercised in the interpretation of the estimated probabilities of these ancestral states.

Figure 6

(a and b) Reconstruction of ancestral states (distribution and tusk form) based on maximum likelihood. Terminal nodes are labeled with species codes as per Figure 2. Distribution coded as per “relaxed” option in Table 5. See Table 3 for details of tusk coding options.

Figure 6

(a and b) Reconstruction of ancestral states (distribution and tusk form) based on maximum likelihood. Terminal nodes are labeled with species codes as per Figure 2. Distribution coded as per “relaxed” option in Table 5. See Table 3 for details of tusk coding options.

Although predictions from H4 do hold in the North Atlantic for at least some tusk coding options, this was not the case for other ocean basins where a greater variety of species co-occur (Table 6). For example, under H1-A, two pairs of species have similar tusks in the North Pacific (state 3—M. stejnegeri, M. carlhubbsi; state 5—M. densirostris, M. peruvianus), whereas in the Southern Hemisphere, three pairs of species have similar tusks (state 1—M. mirus, M. hectori; state 2—M. grayi, M. ginkgodens; state 5—M. densirostris, M. peruvianus). A similar pattern was found with other tusk coding options—namely, that species with tusks that are similar based on some measures, though they may differ based on other measures, do in fact have overlapping distributions in many cases. This finding also broadly contradicts predictions from H3. However, sympatric sister species nonetheless appear to represent a relatively good match with these predictions. Three sister-species pairs overlap in distribution: M. europaeus and M. mirus in the North Atlantic and M. perrini and M. peruvianus and M. stejnegeri and M. densirostris, in the North Pacific. The first two possess divergent tusks under the majority of coding options, and although the tusks of M. stejnegeri and M. densirostris do appear similar based on evaluation of their position in the jaw, size, width, and overall shape (H1-B to-E, H1-G, H1-H), they differ markedly in their length, angle of incline, and whether or not they are raised up on an arch (H1-F, H1-I, H1-J).

Table 6

Tusk forms by ocean basin. Species codes follow Figure 2. Species distributions are as indicated in Figure 5. See Table 3 for details of H1-A to H1-J tusk form coding options. Sister species found in the same ocean basin are indicated by “+” between species codes and character states. M. stejnegeri (Mst) and M. densirostris (Mde—all oceans) are also sister species and co-occur in the North Pacific. M. carlhubbsi (Mca) and M. bowdoini (Mbow) are sister species but do not overlap in distribution.

 Distribution   

 
 North Atlantic North Pacific Southern Hemisphere All oceans 
Species Mbi, Meu + Mmi Mca, Mgin, Mpi + Mpe, Mst Mbow, Mgr, Mhe, Mlay, Mmi, Mpe Mde 
H1-A 4, 2 + 1 3, 2, 1 + 5, 3 3, 2, 1, 4, 1, 5 
H1-B 3, 2 + 1 2, 2, 1 + 2, 2 2, 2, 1, 3, 1, 2 
H1-C 3, 2 + 1 3, 4, 1 + 4, 4 3, 2, 1, 3, 1, 4 
H1-D 4, 2 + 1 3, 3, 1 + 3, 3 3, 3, 1, 4, 1, 3 
H1-E 2, 2 + 2 3, 3, 2 + 1, 3 3, 2, 2, 4, 2, 1 
H1-F 1, 1 + 1 2, 1, 1 + 1, 2 2, 1, 1, 3, 1, 1 
H1-G 1, 2 + 1 3, 4, 1 + 1, 3 3, 3, 1, 2, 1, 1 
H1-H 2, 2 + 1 3, 3, 2 + 1, 3 3, 3, 2, 4, 1, 1 
H1-I 4, 2 + 1 2, 2, 2 + 1, 3 3, 2, 1, 4, 1, 1 
H1-J 1, 1 + 1 1, 1, 1 + 2, 1 1, 1, 1, 1, 1, 2 
 Distribution   

 
 North Atlantic North Pacific Southern Hemisphere All oceans 
Species Mbi, Meu + Mmi Mca, Mgin, Mpi + Mpe, Mst Mbow, Mgr, Mhe, Mlay, Mmi, Mpe Mde 
H1-A 4, 2 + 1 3, 2, 1 + 5, 3 3, 2, 1, 4, 1, 5 
H1-B 3, 2 + 1 2, 2, 1 + 2, 2 2, 2, 1, 3, 1, 2 
H1-C 3, 2 + 1 3, 4, 1 + 4, 4 3, 2, 1, 3, 1, 4 
H1-D 4, 2 + 1 3, 3, 1 + 3, 3 3, 3, 1, 4, 1, 3 
H1-E 2, 2 + 2 3, 3, 2 + 1, 3 3, 2, 2, 4, 2, 1 
H1-F 1, 1 + 1 2, 1, 1 + 1, 2 2, 1, 1, 3, 1, 1 
H1-G 1, 2 + 1 3, 4, 1 + 1, 3 3, 3, 1, 2, 1, 1 
H1-H 2, 2 + 1 3, 3, 2 + 1, 3 3, 3, 2, 4, 1, 1 
H1-I 4, 2 + 1 2, 2, 2 + 1, 3 3, 2, 1, 4, 1, 1 
H1-J 1, 1 + 1 1, 1, 1 + 2, 1 1, 1, 1, 1, 1, 2 

Estimated Divergence Dates

Substitution patterns in our data set did not follow a strict molecular clock model (P < 0.01). Estimates generated using a relaxed molecular clock were in agreement on the order and tempo at which the lineages diverged but differed in their estimates of divergence dates, as expected under the different parameter options. We considered option B2 (emergence of Mesoplodon between 25 and 15 Ma) to be the most likely scenario and only these results are presented here (Fig. 5). The radiation of Mesoplodon beaked whales does not appear to have occurred in a single burst. Instead, new lineages have continued to arise throughout the evolutionary history of this genus. As a result, extant lineages represent a wide range of ages. The youngest lineages include four sister-species pairs, for which divergence estimates range from 10.4 to 5.3 Ma. Assuming current distributions are an accurate reflection of past distributions, the North Atlantic was first colonized by M. bidens, followed later by members of the True's beaked whale clade (three species), whereas M. densirostris (Perrin's beaked whale clade) was a relative latecomer to these waters. The North Pacific was first colonized by two lineages from the straptooth and True's beaked whale clades but subsequently exploited more widely by several members of the Perrin's beaked whale clade (four species). The Southern Hemisphere is largely the domain of the straptooth and Perrin's beaked whale clades, but two members of the True's beaked whale clade have also made successful incursions into this region. Note that although M. densirostris is one of the youngest members of this group, it has the widest distribution and is found in all three ocean basins.

Discussion

Radiation of the Genus Mesoplodon

Of the four hypotheses put forward to explain patterns of Mesoplodon tusk morphology and/or the processes that have driven the diversification of this genus, only two seem to warrant further consideration—sexual selection (H3; Dalebout, 2002) and species recognition (H4; MacLeod, 2000). The co-occurrence of species with similar tusks in the North Pacific and Southern Hemisphere appears to contradict predictions from the latter. However, our coding of the tusks based on various characteristics such as position in the jaw, size, shape, and angle of inclination is somewhat artificial. In reality, the tusks of every species are sufficiently unique that they can be used as diagnostic characters for species identification by researchers (e.g., Allen et al., 2001; Allen, 2007). Further, it is also not clear how other whales perceive these tusks. Here we have assumed that the tusks are a visual cue. Given the echolocation abilities of beaked whales (Johnson et al., 2004), the tusks could also provide a unique acoustic profile. Toothed whales use the phonic lips in their nasal passages to produce sounds that are transmitted through the waxy melon in the forehead. Echoes from objects such as prey or other whales are then received via the lower jaws (Jones, 2005). The tusks are associated with the receiving components of this system, which in whales is quite separate from the sound-producing component. Therefore, it seems unlikely that they function as modifiers of acoustic signals, as would be required if they were involved in the production of species-specific courtship calls.

The species-recognition hypothesis also assumes that there is female choice in the selection of mates and a high potential for mistakes. Ornaments that serve as cues for species recognition are usually possessed by both sexes (Andersson, 1994). Where ornaments are possessed by only one sex, mate choice is the prerogative of the other sex (West-Eberhard, 1983). Given that it is the adult males who have tusks, female choice is expected to be the dominant pattern. Sex-specific recognition cues can help animals avoid mistakes in breeding with similar-looking species in crowded ecosystems, or, as may be the case for Mesoplodon beaked whales, if encounters with conspecifics are rare. If mistakes in mate identification did occur, we would expect to occasionally encounter hybrids. In baleen whales, hybrids between blue and fin whales (Balaenoptera musculus and B. physalus) have been encountered on several occasions, whereas in toothed whales, hybridization can occur among several dolphin species, though this generally happens only in captivity (see review by Bérubé and Aguilar, 1998). However, no hybrid beaked whales have been reported to date from morphology or genetics. Overall, the species recognition hypothesis does not therefore appear to offer sufficient explanation for the diversity and distribution of male tusk form in Mesoplodon beaked whales, though selection for species recognition cues could nonetheless help maintain divergent tusk morphologies among some sympatric species.

None of the three main Mesoplodon clades was confined to a single ocean basin, as might be expected from strict sympatric speciation driven by sexual selection. However, three of the four sister-species pairs have both overlapping distributions and divergent tusk forms, as predicted by this hypothesis. The only sister-species pair that does not follow this pattern occurs in different ocean basins (Fig. 5). These sister-species pairs are among the youngest lineages in this genus. This pattern, as revealed by our analyses, suggests that although sexual selection may have played an important role in the diversification of this genus, the broad evolutionary time scale on which this process likely occurred, together with the high dispersal abilities of these species and the plasticity of tusk form, may have overwritten much of the evidence. Species evolving in sympatry through sexual selection could subsequently disperse to new areas where they would incur new costs for the effective exploitation of new niches. Little is known about the diet of these whales beyond their preference for deep-water squid (MacLeod et al., 2003). Although the tusks are not used for feeding, large tusks are likely to be costly to grow and maintain. In some species, the tusks may even hinder feeding. For example, the long strap-like teeth of M. layardii grow to curve over the upper jaw and limit the gape to only a few centimeters (Reeves et al., 2002). More importantly perhaps, violent male-male combat is also a costly activity that can result in injury. Although there is no doubt that the original function of the male tusks is as weapons, it is not clear that all extant Mesoplodon species engage in violent combat. The widespread occurrence of these encounters has been inferred from the heavy, highly visible scarring accumulated by adult males, but this behavior has never been observed (Mead et al., 1982; Heyning, 1984). Note that it is also possible that this scarring is used by females as an indicator of male quality (Macleod, 1998). Insufficient fresh specimens have been examined for some Mesoplodon species to determine whether they all engage in equally violent contests. In some cases, it is possible that male-male combat no longer occurs (e.g., M. ginkgodens; Reeves et al., 2002), and as a result, selection on the tusks for use as weapons may have been relaxed.

Tempo of Speciation

Due to the paucity of beaked whale fossils of known age and with clear links to extant lineages, our estimates of divergence dates were reliant on a single calibration point. The window of time covered by this calibration point, and by the different parameter scenarios explored, was very wide (30 to 5 Ma). As a result, our divergence date estimates for Mesoplodon species have large standard errors and should be interpreted with caution. As new fossil beaked whales come to light and their affinities to modern taxa are clarified (Lambert, 2005; Lambert and Louwye, 2006), these estimates will likely require revision. It is worth noting, however, that our estimate for the divergence of the youngest sister-species pair in this genus (M. perrini and M. peruvianus) at 5.3 Ma is concordant with estimates from mtDNA based on the general mammalian rate of approximately 1% divergence per million years (Dalebout, 2002).

However, investigation of the relative tempo at which speciation occurred is not reliant on robust fossil calibrations. Speciation has clearly continued throughout the evolution of the genus, no matter to what date its initial emergence is constrained. This is an interesting observation in itself as radiations driven primarily by sexual selection often seem to occur in a quick burst over a short period of time—e.g., cichlids in the African Great Lakes (Galis, 1998; Genner et al., 2007). This suggests that the radiation of Mesoplodon beaked whales has perhaps not been as rapid as would expected if it were driven by strong, sustained sexual selection.

The potential impact of missing lineages should, however, also be considered. By necessity, we have been able to assess only the diversity and distribution of extant species. Many other Mesoplodon species may have fallen by the wayside during the evolution of this group and likely also influenced the tempo and mode of this radiation. Ghost-lineage analysis could provide some insight into the proportion of missing taxa (Teeling et al., 2005) but unfortunately is reliant on multiple fossil calibrations, which are not available in this case.

Agreement between Molecules and Morphology

Here we have presented the first robust, highly resolved molecular phylogeny for Mesoplodon beaked whales, one of the most speciose, yet least well-known, of cetacean genera. This work, based on nuclear intron sequences, complements and supports earlier studies in which mtDNA data were used to confirm the genetic distinctiveness of all previously recognized species in this group (Dalebout et al., 2004) and describe two additional species (Dalebout et al., 2002; van Helden et al., 2002). To date, there have been no explicit attempts to reconstruct evolutionary relationships among beaked whales based on morphological features. There are several reasons for this. First, many species in this family are rare, with several known from fewer than 30 individuals (Reeves et al., 2002). Second, most morphological work on cetaceans focuses on cranial characters. In beaked whales, these features can differ strongly by sex and age class, further reducing the number of specimens available for inter-species comparison (e.g., Mead, 1989). Third, many species have wide geographic distributions. As a result, specimens are scattered in museum collections throughout the world. Finally, cranial morphology in this group appears to be relatively conserved across species, perhaps due to ecological constraints associated with deep-diving. Attempts to clarify the morphological distinctiveness of these species have traditionally placed most emphasis on the male tusks and features of the “vertex,” the elevated crest formed by the cranial bones behind the superior nares (e.g., Moore, 1966, 1968). From a phylogenetic perspective, it is not clear that small-scale differences in the arrangement of the vertex bones have any real adaptive function and, as such, can serve as a useful character to reconstruct evolutionary relationships in this group.

Morphological assessments have, however, suggested close relationships between several Mesoplodon species that this study has subsequently confirmed to be sister-species pairs. For example, Moore (1957, 1960) proposed a sister-species relationship between M. mirus and M. europeaus based on cranial morphology, whereas M. carlhubbsi and M. bowdoini were initially described as the same species (Hubbs, 1946) before further morphological analyses revealed their distinctiveness (Moore, 1963). Interestingly, although these sister-species relationships were supported by our nuclear intron work, mtDNA analyses failed to resolve the close relationship between the latter pair (although it was not strongly excluded either; Dalebout et al., 1998, 2004). This may be due to the relatively old age of this divergence (approximately 10.4 Ma; Fig. 5) and the accumulation of phylogenetic noise (homoplasies) by the mtDNA genome since this time. The other sister-species pairs identified here are all younger (approximately ≤ 7.4 Ma) and were also resolved by previous mtDNA analyses (although like most other higher-level nodes, such groupings generally received low BS scores in mtDNA reconstructions; e.g., Dalebout et al., 2002). Cranial morphology has also suggested a link between M. bowdoini and M. traversii (Reyes et al., 1996; Baker, 2001), the only described Mesoplodon species not included in this study. M. traversii is currently known from three partial skulls from the Southern Hemisphere and its external appearance remains unknown (van Helden et al., 2002). mtDNA analyses suggest that M. traversii is likely a member of the straptooth clade and of similar age as the other three species (Dalebout, 2002). The position of the tusks in M. traversii, which overlap with the posterior end of the symphysis, also supports this conclusion. Confirmation of this relationship will require nuclear data and fresh tissue samples from which high-quality DNA can be extracted. Intriguingly, tusk form in M. traversii is similar to that of M. layardii (van Helden et al., 2002) even though their skulls differ, such that M. traversii appears to be a morphological composite of the other taxa in this clade.

Although cranial features can be used to infer sister-species relationships in this family, similarity in tusk form can be misleading given the plasticity of this feature as demonstrated by our analyses (Fig. 6a, Fig. 6b). The large apical tusks of both M. hectori and M. perrini initially led researchers to assume that they represented a single species (Mead, 1981; Mead and Baker, 1987). When mtDNA analyses revealed the distinctiveness and non-sister-species relationship of these rare whales (Dalebout et al., 1998, 2002), the morphological features were re-evaluated. As expected from the deep mtDNA divergence (Dalebout et al., 2002)—a pattern later confirmed by analysis of a nuclear intron (Dalebout et al., 2004)—there are several fixed morphological differences between these taxa, although they are relatively subtle (Dalebout et al., 2002). The phylogeny presented here provides further confirmation of the distinctiveness of these species (Fig. 5) and suggests that their similar tusks are the result of parallel evolution and the retention or re-emergence of ancestral character states. It is worth noting that although M. hectori and M. perrini resemble one another in their conservative vertex morphology (Mead and Baker, 1987), this resemblance is stronger still between M. perrini and its real sister species, M. peruvianus, though their tusks are highly divergent.

Conclusions and Further Work

We have tested predictions from four hypotheses regarding the evolution and diversification of a speciose genus of whales using a robust, highly resolved molecular phylogeny. The patterns observed in three of the four sister-species pairs identified strongly suggest that sexual selection on weaponry in the form of the male tusks, together perhaps with selection for species recognition cues, has played an important role in this unique radiation. To our knowledge, this is the first time that sexual selection has been explicitly implicated in the radiation of a mammalian group outside terrestrial ungulates. However, the extent to which sexual selection has occurred, and the effect of other selection pressures (e.g., female mate choice, niche separation, and diet), is difficult to ascertain due to our lack of knowledge about many aspects of beaked whale biology. Insights into social organization, including the role of male-male combat and female mate choice, will be difficult to obtain as most Mesoplodon species are rarely encountered at sea and generally cannot be observed for extended periods (Reeves et al., 2002). However, sightings of single adult males accompanying multiple adult females suggest that these whales may have a polygynous mating system (McSweeney et al., 2007), a system that often does not allow for significant female choice. Dead stranded or beachcast animals could also provide a wealth of information about niche separation if the stomach contents are collected (MacLeod et al., 2003). This should be done far more regularly than is currently the case in many areas.

Robust reconstruction of relationships among the other genera in this family will require data from a suite of nuclear introns similar to those analyzed here for Mesoplodon beaked whales. Previous analyses based on two introns (ACT and DQA) failed to resolve these relationships due to the presence of very short inter-node lengths at the base of the tree, suggestive of a rapid radiation event (Dalebout, 2002). Closely related outgroups, which could help determine the polarity of character state changes, are also unavailable. Beaked whales are among the oldest of cetacean lineages (Barnes et al., 1985). The closest sister groups are the Platanistidae (river dolphins) and Physeteridae (sperm whales; Heyning, 1997; Cassens et al., 2000), both of which have been reduced to essentially a single extant species (a stark contrast to the Ziphiidae). The long branches of these sister groups may therefore be as much of a hindrance as a help in the reconstruction of a robust phylogeny (Felsenstein, 1978). As a result, nuclear data, including short interspersed repeated elements (SINES; Nikaido et al., 2001), from a range of cetacean families will likely be needed to resolve inter-generic relationships among beaked whales.

Acknowledgment

For collection and access to samples and specimens, we thank New Zealand Department of Conservation field center staff (NZ); Kelly Robertson, US NMFS Southwest Fisheries Science Center (SW); Bob Reid, Scottish Agricultural College (SAC); Nick Gales, Australian Antarctic Division (WA); Marty Haulena, the Marine Mammal Center (TMMC), California; Paul Jepson, Institute of Zoology, London (NHMUK); John Heyning, Los Angeles County Museum of Natural History (LAM); Charley Potter, Smithsonian Institution US National Museum of Natural History (NE, WAM); Cath Kemper, South Australian Museum (SAM); Tadasu Yamada, National Science Museum, Japan (TSM); and John Wang, FormosaCetus, Taiwan (TW). Abbreviations in parentheses refer to species codes in Appendix 1. We thank Sarah Mesnick, US NMFS Southwest Fisheries Science Center, for initiating discussion regarding the role of sexual selection in the radiation of the Ziphiidae; Robin Beck, UNSW, for assistance with phylogenetic analyses; Vivian Ward, University of Auckland for beaked whale jaw images; and Andrew Glaser, New Zealand Department of Conservation, and Gina Lento, University of Auckland, for use of Gray's beaked whale photographs. This article benefitted from comments by William B. Sherwin, Systematic Biology Editor-in-Chief Jack Sullivan, Associate Editor Michael Charleston, and an anonymous reviewer. Funding for DNA sequencing was provided by grants to C.S.B. from the New Zealand Marsden Fund, the US Marine Mammal Commission, and the Committee for Research and Exploration of the National Geographic Society. M.L.D. is supported by a UNSW Vice-Chancellor's Postdoctoral Fellowship.

References

Akaike
H.
Information theory and an extension of the maximum likelihood principle
Second International Symposium on Information Theory
 , 
1973
Budapest
Akademiai Kiado
(pg. 
267
-
281
)
Allen
B. M.
Smithsonian Institution Marine Mammal Program beaked whale identification guide
 , 
2007
 
Allen
B. M.
Chapman
R. E.
Mead
J. G.
Species and sex determination of Mesoplodon, Ziphius based on mandibular measurements
2001
Abstract, 14th Biennial Conference on the Biology of Marine Mammals
Vancouver, BCCanada
 
Available from the Society for Marine Mammalogy, PO Box 692042, Orlando, FL 32869–2042, USA
Andersson
M.
Sexual selection
 , 
1994
Princeton, New Jersey
Princeton University Press
Arnason
U.
Gullberg
A.
Cytochrome b nucleotide sequences and the identification of five primary lineages of extant cetaceans
Mol. Biol. Evol.
 , 
1996
, vol. 
13
 (pg. 
407
-
417
)
Auffray
C.
Lillie
J. W.
Korman
A. J.
Boss
J. M.
Frechin
N.
Guillemnot
F.
Cooper
J.
Mulligan
R. C.
Strominger
J. L.
Structure and expression of HLA-DQA and-DXA genes: Interallelic alternate splicing of the HLA-DQA gene and functional splicing of the HLA-DQX gene using a retroviral vector
Immunogenetics
 , 
1987
, vol. 
26
 (pg. 
63
-
73
)
Baker
A. N.
Status, relationships and distribution of Mesoplodon bowdoini Andrews, 1908 (Cetacea: Ziphiidae)
Mar. Mammal. Sci.
 , 
2001
, vol. 
17
 (pg. 
473
-
493
)
Baker
C. S.
Slade
R. W.
Bannister
J. L.
Abernethy
R. B.
Weinrich
M. T.
Lien
J.
Urban
J.
Corkeron
P.
Calmabokidis
J.
Vasquez
O.
Palumbi
S. R.
Hierarchical structure of mitochondrial DNA gene flow among humpback whales Megaptera novaeangliae, world-wide
Mol. Ecol.
 , 
1994
, vol. 
3
 (pg. 
313
-
327
)
Barlow
J.
Ferguson
M.
Perrin
W. F.
Ballance
L.
Gerrodette
T.
Joyce
G. G.
MacLeod
C. D.
Mullin
K.
Palka
D. L.
Waring
G. T.
Abundance and densities of beaked and bottlenose whales (family Ziphiidae)
J. Cet. Res. Man.
 , 
2006
, vol. 
7
 (pg. 
263
-
270
)
Barnes
L. G.
Domning
D. P.
Ray
C. E.
Status of studies on fossil marine mammals
Mar. Mammal. Sci.
 , 
1985
, vol. 
1
 (pg. 
15
-
53
)
Berglund
A.
Bisazza
A.
Pilastro
A.
Armaments and ornaments: An evolutionary explanation of traits of dual utility
Biol. J. Linn. Soc.
 , 
1996
, vol. 
58
 (pg. 
385
-
399
)
Bérubé
M.
Aguilar
A.
A new hybrid between a blue whale, Balaenoptera musculus, and a fin whale, B. physalus: Frequency and implications of hybridization
Mar. Mammal. Sci.
 , 
1998
, vol. 
14
 (pg. 
82
-
98
)
Carwardine
M.
Whales, dolphins and porpoises
 , 
1995
London
HarperCollins
Cassens
I.
Vicario
S.
Waddell
V. G.
Balchowsky
H.
Van Belle
D.
Wang
D.
Chen
F.
Lal Mohan
R. S.
Simões-Lopes
P. C.
Bastida
R.
Meyer
A.
Stanhope
M. J.
Milinkovitch
M. C.
Independent adaptation to riverine habitats allowed survival of ancient cetacean lineages
Proc. Natl. Acad. Sci. USA
 , 
2000
, vol. 
97
 (pg. 
11343
-
11347
)
Dalebout
M. L.
Species identity, genetic diversity and molecular systematic relationships among the Ziphiidae (beaked whales)
 , 
2002
Auckland, New Zealand
School of Biological Sciences, University of Auckland
 
PhD thesis
Dalebout
M. L.
Baker
C. S.
Cockcroft
V. G.
Mead
J. G.
Yamada
T. K.
A comprehensive and validated molecular taxonomy of beaked whales, family Ziphiidae
J. Hered.
 , 
2004
, vol. 
95
 (pg. 
459
-
473
)
Dalebout
M. L.
Mead
J. G.
Baker
C. S.
Baker
A. N.
van Helden
A. L.
A new species of beaked whale Mesoplodon perrini sp. n. (Cetacea: Ziphiidae) discovered through phylogenetic analyses of mitochondrial DNA sequences
Mar. Mammal. Sci.
 , 
2002
, vol. 
18
 (pg. 
577
-
608
)
Dalebout
M. L.
Ross
G. J. B.
Baker
C. S.
Anderson
R. C.
Best
P. B.
Cockcroft
V. G.
Hinsz
H. L.
Peddemors
V.
Pitman
R. L.
Appearance, distribution and genetic distinctiveness of Longman's beaked whale, Indopacetus pacificus
Mar. Mammal. Sci.
 , 
2003
, vol. 
19
 (pg. 
421
-
461
)
Dalebout
M. L.
van Helden
A.
Van Waerebeek
K.
Baker
C. S.
Molecular genetic identification of southern hemisphere beaked whales (Cetacea: Ziphiidae)
Mol. Ecol.
 , 
1998
, vol. 
7
 (pg. 
687
-
694
)
de Queiroz
K.
For consensus (sometimes)
Syst. Biol.
 , 
1993
, vol. 
42
 (pg. 
368
-
372
)
Eberhard
W. G.
Horned beetles
Sci. Am.
 , 
1980
, vol. 
242
 (pg. 
166
-
182
)
Emlen
D. J.
Marangelo
J.
Ball
B.
Cunningham
C. W.
Diversity in the weapons of sexual selection: Horn evolution in the beetle genus Onthophagus (Coleoptera: Scarabaeidae)
Evolution
 , 
2005
, vol. 
59
 (pg. 
1060
-
1084
)
Farris
J. S.
Kallersjo
M.
Kluge
A. G.
Bult
C.
Testing the significance of incongruence
Cladistics
 , 
1995
, vol. 
10
 (pg. 
315
-
319
)
Felsenstein
J.
Cases in which parsimony and compatibility methods will be positively misleading
Syst. Zool.
 , 
1978
, vol. 
27
 (pg. 
401
-
410
)
Fordyce
R. E.
Perrin
W. F.
Wursig
B.
Thewissen
J. G. M.
Fossil record
Encyclopedia of marine mammals
 , 
2002
San Diego
Academic Press
(pg. 
453
-
471
)
Galis
F.
Why are there so many cichlid species?
TREE
 , 
1998
, vol. 
13
 (pg. 
1
-
2
)
Geist
V.
Kramer
L.
Pliner
P.
Alloway
T.
On weapons, combat and ecology
Aggression, dominance, and individual spacing
 , 
1978
New York
Plenum
(pg. 
1
-
30
)
Genner
M. J.
Seehausen
O.
Lunt
D. H.
Joyce
D. A.
Shaw
P. W.
Carvalho
G. R.
Turner
G. F.
Age of cichlids: New dates for ancient lake fish radiations
Mol. Biol. Evol.
 , 
2007
, vol. 
24
 (pg. 
1269
-
1282
)
Heyning
J. E.
Functional morphology involved in intraspecific fighting of the beaked whale, Mesoplodon carlhubbsi
Can. J. Zool.
 , 
1984
, vol. 
62
 (pg. 
1645
-
1654
)
Heyning
J. E.
Sperm whale phylogeny revisited: Analysis of the morphological evidence
Mar. Mammal. Sci.
 , 
1997
, vol. 
13
 (pg. 
596
-
613
)
Hubbs
C. L.
First records of two beaked whales, Mesoplodon bowdoini Ziphius cavirostris from the Pacific Coast of the United States
J. Mammal.
 , 
1946
, vol. 
27
 (pg. 
242
-
255
)
Huelsenbeck
J. P.
Ronquist
F.
MrBayes: Bayesian inference of phylogenetic trees
Bioinformatics
 , 
2001
, vol. 
17
 (pg. 
754
-
755
)
Johnson
M.
Madsen
P.
Zimmer
W.
de Soto
N.
Tyack
P.
Beaked whales echolocate on prey
Proc. R. Soc. Lond. B
 , 
2004
, vol. 
271
 (pg. 
S383
-
S386
)
Jones
G.
Echolocation
Curr. Biol.
 , 
2005
, vol. 
15
 (pg. 
R484
-
R488
)
Kaneshiro
K. Y.
Sexual selection and direction of evolution in the biosystematics of Hawaiian Drosophilidae
Ann. Rev. Entomol.
 , 
1983
, vol. 
28
 (pg. 
161
-
178
)
Kellogg
R.
The history of whales-their adaptation to life in water (Part I)
Quart. Rev. Biol.
 , 
1928
, vol. 
3
 (pg. 
29
-
76
)
Kellogg
R.
The history of whales-their adaptation to life in water (Part II)
Quart. Rev. Biol.
 , 
1928
, vol. 
3
 (pg. 
174
-
208
)
Kishino
H.
Hasegawa
M.
Evaluation of the maximum likelihood estimate of the evolutionary tree topologies from DNA sequence data, and branching order in Hominoidea
Mol. Evol.
 , 
1989
, vol. 
29
 (pg. 
170
-
179
)
Kishino
H.
Thorne
J. L.
Bruno
W. J.
Performance of a divergence time estimation method under a probabilistic model of rate evolution
Mol. Biol. Evol.
 , 
2001
, vol. 
18
 (pg. 
352
-
361
)
Kumar
S.
Tamura
K.
Nei
M.
MEGA3: Integrated software for molecular evolutionary genetics analysis and sequence alignment
Brief. Bioinf.
 , 
2004
, vol. 
5
 (pg. 
150
-
163
)
Lambert
O.
Systematics and phylogeny of the fossil beaked whales Ziphirostrum du Bus, 1868 and Choneziphius Duvernoy, 1851 (Mammalia, Cetacea, Odontoceti), from the Neogene of Antwerp (North of Belgium)
Geodiversitas
 , 
2005
, vol. 
27
 (pg. 
443
-
497
)
Lambert
O.
Louwye
S.
Archeoziphius microglenoideus, a new primitive beaked whale (Mammalia, Cetacea, Odontoceti) from the Middle Miocene of Belgium
J. Vert. Paleo.
 , 
2006
, vol. 
26
 (pg. 
182
-
191
)
Lundrigan
B.
Morphology of horns and fighting behaviour in the family Bovidae
J. Mammal.
 , 
1996
, vol. 
77
 (pg. 
462
-
475
)
Lyons
L. A.
Laughlin
T. F.
Copeland
N. G.
Jenkins
N. A.
Womack
J. E.
O'Brien
S. J.
Comparative anchor tagged sequences (CATS) for integrative mapping of mammalian genomes
Nat. Genet.
 , 
1997
, vol. 
15
 (pg. 
47
-
56
)
MacLeod
C. D.
Intraspecific scarring in odontocete cetaceans: An indicator of male “quality” in aggressive social interactions?
J. Zool. Lond.
 , 
1998
, vol. 
244
 (pg. 
71
-
77
)
MacLeod
C. D.
Species recognition as a possible function for variations in position and shape of the sexually dimorphic tusks of Mesoplodon beaked whales
Evolution
 , 
2000
, vol. 
54
 (pg. 
2171
-
2173
)
MacLeod
C. D.
Perrin
W. F.
Pitman
R. L.
Barlow
J.
Ballance
L.
D'Amico
A.
Gerrodette
T.
Joyce
G.
Mullin
K. D.
Palka
D. L.
Waring
G. T.
Known and inferred distributions of beaked whale species (Ziphiidae: Cetacea)
J. Cet. Res. Man.
 , 
2006
, vol. 
7
 (pg. 
271
-
286
)
MacLeod
C. D.
Santos
M. B.
Pierce
G. J.
Review of data on diets of beaked whales: Evidence of niche separation and geographic segregation
J. Mar. Biol. Assoc. (UK)
 , 
2003
, vol. 
83
 (pg. 
651
-
665
)
Maddison
W. P.
Maddison
D. R.
MacClade: Analysis of phylogeny and character evolution
 , 
2005
Sunderland, Massachusetts
Sinauer Associates
Maddison
W. P.
Maddison
D. R.
Mesquite: A modular system for evolutionary analysis. Version 1.12
 , 
2006
 
May-Collado
L.
Agnarsson
I.
Cytochrome b and Bayesian inference of whale phylogeny
Mol. Phylog. Evol.
 , 
2006
, vol. 
38
 (pg. 
344
-
354
)
McSweeney
D. J.
Baird
R. W.
Mahaffy
S. D.
Site fidelity, associations, and movements of Cuvier's (Ziphius cavirostris) and Blainville's (Mesoplodon densirostris) beaked whales off the island of Hawai'i
Mar. Mamm. Sci.
 , 
2007
, vol. 
23
 (pg. 
666
-
687
)
Mead
J. G.
First records of Mesoplodon hectori (Ziphiidae) from the Northern Hemisphere and a description of the adult male
J. Mammal.
 , 
1981
, vol. 
62
 (pg. 
430
-
432
)
Mead
J. G.
Ridgway
S. H.
Harrison
R.
Beaked whales of the genus Mesoplodon
Handbook of marine mammals
 , 
1989
London
Academic Press
(pg. 
349
-
430
)
Mead
J. G.
Baker
A. N.
Notes on the rare beaked whale, Mesoplodon hectori (Gray)
J. R. Soc. N. Z.
 , 
1987
, vol. 
17
 (pg. 
303
-
312
)
Mead
J. G.
Walker
W. A.
Houck
W. J.
Biological observations on Mesoplodon carlhubbsi (Cetacea: Ziphiidae)
Smiths. Contrib. Zool.
 , 
1982
, vol. 
344
 (pg. 
1
-
25
)
Messenger
S. L.
McGuire
J. A.
Morphology, molecules and the phylogenetics of cetaceans
Syst. Biol.
 , 
1998
, vol. 
47
 (pg. 
90
-
124
)
Moore
J. C.
Differences between the beaked whales Mesoplodon mirus Mesoplodon gervaisi
Am. Mus. Nov.
 , 
1957
, vol. 
1831
 (pg. 
1
-
25
)
Moore
J. C.
New records of the gulf-stream beaked whale, Mesoplodon gervaisi, and some taxonomic considerations
Am. Mus. Nov.
 , 
1960
, vol. 
1993
 (pg. 
1
-
35
)
Moore
J. C.
Recognizing certain species of beaked whales of the Pacific Ocean
Am. Mid. Nat.
 , 
1963
, vol. 
70
 (pg. 
396
-
428
)
Moore
J. C.
Diagnoses and distributions of beaked whales of the genus Mesoplodon known from North American waters in whales, dolphins and porpoises
International Symposium on Cetacean Research
 , 
1966
Washington, DC
University of California Press
(pg. 
32
-
16
)
Moore
J. C.
Relationships among the living genera of beaked whales
Field. Zool.
 , 
1968
, vol. 
53
 (pg. 
209
-
298
)
Nikaido
M.
Matsuno
F.
Hamilton
H.
Brownell
R. L. J.
Cao
Y.
Ding
W.
Zuoyan
Z.
Shedlock
A. M.
Fordyce
R. E.
Hasegawa
M.
Okada
N.
Retroposon analysis of major cetacean lineages: The monophyly of toothed whales and the paraphyly of river dolphins
Proc. Natl. Acad. Sci. USA
 , 
2001
, vol. 
98
 (pg. 
7384
-
7389
)
Owen
R.
Monograph on the British fossil Cetacea from the Red Crag
 , 
1870
London
The Palaeontological Society
Pagel
M.
Meade
A.
Bayesian analysis of correlated evolution of discrete characters by reversible-jump Markov Chain Monte Carlo
Am. Nat.
 , 
2006
, vol. 
167
 (pg. 
808
-
825
)
Palumbi
S. R.
Baker
C. S.
Contrasting population structure from nuclear intron sequences and mtDNA of humpback whales
Mol. Biol. Evol.
 , 
1994
, vol. 
11
 (pg. 
426
-
435
)
Posada
D.
ModelTest Server: A web-based tool for the statistical selection of models of nucleotide substitutions online
Nucleic Acids Res.
 , 
2006
, vol. 
34
 (pg. 
W700
-
W703
)
Posada
D.
Buckley
T. R.
Model selection and model averaging in phylogenetics: Advantages of Akaike Information Criterion and Bayesian approaches over likelihood ratio tests
Syst. Biol.
 , 
2004
, vol. 
53
 (pg. 
793
-
808
)
Posada
D.
Crandall
K. A.
ModelTest: Testing the model of DNA substitution
Bioinformatics Appl. Notes
 , 
1998
, vol. 
14
 (pg. 
817
-
818
)
Rannala
B.
Yang
Z.
Probability distribution of molecular evolutionary trees: A new method of phylogenetic inference
J. Mol. Evol.
 , 
1996
, vol. 
43
 (pg. 
304
-
311
)
Reeves
R. R.
Brent
S. S.
Clapham
P. J.
Powell
J. A.
Folkens
P. A.
National Audubon Society guide to marine mammals of the world
 , 
2002
New York
Chanticleer Press
Reyes
J. C.
Van Waerebeek
K.
Cárdenas
J. C.
Yañez
J. L.
Mesoplodon bahamondi sp. n. (Cetacea, Ziphiidae), a living beaked whale from the Juan Fernandez Archipelago, Chile
Bol. Mus. Nac. Hist. Nat. Chile
 , 
1996
, vol. 
45
 (pg. 
31
-
44
)
Ronquist
F.
Huelsenbeck
J. P.
MrBayes 3: Bayesian phylogenetic inference under mixed models
Bioinformatics
 , 
2003
, vol. 
19
 (pg. 
1572
-
1574
)
Ronquist
F.
Huelsenbeck
J. P.
van der Mark
P.
MrBayes 3.1 Manual
 , 
2005
 
Rutschmann
F.
Bayesian molecular dating using PAML/ multidivtime. A step-by-step manual
 , 
2005
Switzerland
University of Zurich
 
Available at http://www.plant.ch
Sambrook
E.
Fritsch
F.
Maniatis
T.
Molecular cloning
 , 
1989
Cold Spring Harbor, New York
Cold Spring Harbor Press
Sasaki
T.
Nikaido
M.
Wada
S.
Yamada
T. K.
Cao
Y.
Hasegawa
M.
Okada
N.
Balaenoptera omurai is a newly discovered baleen whale that represents an ancient evolutionary lineage
Mol. Phylogenet. Evol.
 , 
2006
, vol. 
41
 (pg. 
40
-
52
)
Seehausen
O.
van Alphen
J. M.
Can sympatric speciation by disruptive sexual selection explain rapid evolution of cichlid diversity in Lake Victoria?
Ecol. Lett.
 , 
1999
, vol. 
2
 (pg. 
262
-
271
)
Sekiguchi
K.
Klages
N. T. W.
Best
P. B.
The diet of the strap-toothed whales (Mesoplodon layardii)
J. Zool. Lond.
 , 
1996
, vol. 
239
 (pg. 
453
-
463
)
Shimodaira
H.
An approximately unbiased test of phylogenetic tree selection
Syst. Biol.
 , 
2002
, vol. 
51
 (pg. 
492
-
508
)
Shimodaira
H.
Hasegawa
M.
Multiple comparisons of log-likelihoods with applications to phylogenetic inference
Mol. Biol. Evol.
 , 
1999
, vol. 
16
 (pg. 
1114
-
1116
)
Shimodaira
H.
Hasegawa
M.
CONSEL: For assessing the confidence of phylogenetic tree selection
Bioinformatics
 , 
2001
, vol. 
17
 (pg. 
1246
-
1247
)
Shine
R.
Sexual selection and sexual dimorphism in the Amphibia
Copeia
 , 
1979
, vol. 
2
 (pg. 
297
-
306
)
Shuster
S.
Wade
M.
Mating systems and strategies
 , 
2003
Princeton, New Jersey
Princeton University Press
Strimmer
K.
Rambaut
A.
Inferring confidence sets of possibly misspecified gene trees
Proc. R. Soc. Lond. B
 , 
2002
, vol. 
269
 (pg. 
137
-
142
)
Swofford
D. L.
PAUP*. Phylogenetic analysis using parsimony (*and other methods). Version 4
 , 
2003
Sunderland, Massachusetts
Sinauer Associates
Teeling
E. C.
Springer
M. S.
Madsen
O.
Bates
P.
O'Brien
S. J.
Murphy
W. J.
A molecular phylogeny for bats illuminates biogeography and the fossil record
Science
 , 
2005
, vol. 
307
 (pg. 
580
-
584
)
Thorne
J. L.
Kishino
H.
Painter
I. S.
Estimating the rate of evolution of the rate of molecular evolution
Mol. Biol. Evol.
 , 
1998
, vol. 
15
 (pg. 
1647
-
1657
)
Turelli
M.
Heritable genetic variation via mutation-selection balance: Lerch's zeta meets the abdominal bristle
Theor. Popul. Biol.
 , 
1984
, vol. 
25
 (pg. 
138
-
193
)
van Helden
A. L.
Baker
A. N.
Dalebout
M. L.
Reyes
J. C.
Van Waerebeek
K.
Baker
C. S.
Resurrection of Mesoplodon traversii (Gray, 1874), senior synonym of M. bahamondi Reyes, Van Waerebeek, Cárdenas and Yañez, 1995 (Cetacea: Ziphiidae)
Mar. Mammal. Sci.
 , 
2002
, vol. 
18
 (pg. 
609
-
621
)
Wada
S.
Oishi
M.
Yamada
T. K.
A newly discovered species of living baleen whale
Nature
 , 
2003
, vol. 
426
 (pg. 
278
-
281
)
West-Eberhard
M. J.
Sexual selection, social competition, and speciation
Quart. Rev. Biol.
 , 
1983
, vol. 
58
 (pg. 
155
-
178
)
Wilson
E. O.
Diversity of Life
 , 
1992
Cambridge, Massachsetts
Harvard University Press
Yang
Z.
PAML: A program package for phylogenetic analysis by maximum likelihood
Comp. Appl. BioSci.
 , 
1997
, vol. 
13
 (pg. 
555
-
556
)

Appendix 1

Identification codes of specific individuals used to generate nuclear intron sequences. Gene names are listed across the top, species names are represented by three-or four-letter codes. Mesoplodon species codes follow Figure 2. Tsh, Tasmacetus shepherdi; Zca, Ziphius cavirostris. Different identification numbers across the same species indicate where chimeric concatenated sequences were used for phylogenetic analysis.

Species BGN CAT RHO CTLA3 CHRNA1 ACT DQA 
Mbi MbiSAC13091 MbiSAC13091 MbiSAC13091 MbiNE3070 MbiSAC13091 MbiWAM490 MbiSAC13091 
Mbow MbowNZ04 MbowNZ04 MbowNZ04 MbowNZ06 MbowNZ04 MbowSAM18047 MbowMhe01 
Mca McaSW73 McaSW1563 McaSW1563 McaSW1563 McaSW73 McaSW73 McaSW1563 
Mde MdeNHMUK MdeNHMUK MdeNZ02 MdeNZ02 MdeNHMUK MdeNHMUK MdeNHMUK 
Meu MeuSW7444 MeuSW7444 MeuSW4120 MeuSW4120 MeuSW7444 MeuSW7443 MeuSW4120 
Mgin MginNZ03 Mgin01TW Mgin01TW MginNZ03 Mgin01TW MginNZ03 MginNZ03 
Mgr MgrNZ67 MgrNZ67 MgrNZ55 MgrNZ55 MgrNZ55 MgrNZ01 MgrNZ02 
Mhe MheWA02 MheNZ02 MheNZ02 MheWA02 MheNZ02 MheWA02 MheMgr09 
Mlay MlayNZ10 MlayNZ08 MlayNZ08 MlayNZ10 MlayNZ10 MlayNZ10 MlaySAM9788 
Mmi MmiSW4968 MmiSW4972 MmiSW4968 MmiSW4972 MmiSW4968 MmiSW4968 MmiSW4968 
Mpi TMMCC75 TMMCC75 TMMCC75 TMMCC75 TMMCC75 TMMCC75 TMMCC75 
Mpe MpeLAM95654 MpeLAM95654 MpeLAM95654 MpeLAM95654 MpeLAM95654 MpeLAM95654 MpeLAM95654 
Mst MstSW4962 MstSW9491 MstSW9491 MstSW9491 MstSW9491 MstTSM30135 MstSW4962 
Zca ZcaNZ12 ZcaNZ12 ZcaNZ12 ZcaNZ12 ZcaNZ06 ZcaNZ06 ZcaNZ12 
Tsh TshNZ01 TshNZ01 TshNZ01 TshNZ01 TshNZ01 TshNZ02 TshNZ01 
Species BGN CAT RHO CTLA3 CHRNA1 ACT DQA 
Mbi MbiSAC13091 MbiSAC13091 MbiSAC13091 MbiNE3070 MbiSAC13091 MbiWAM490 MbiSAC13091 
Mbow MbowNZ04 MbowNZ04 MbowNZ04 MbowNZ06 MbowNZ04 MbowSAM18047 MbowMhe01 
Mca McaSW73 McaSW1563 McaSW1563 McaSW1563 McaSW73 McaSW73 McaSW1563 
Mde MdeNHMUK MdeNHMUK MdeNZ02 MdeNZ02 MdeNHMUK MdeNHMUK MdeNHMUK 
Meu MeuSW7444 MeuSW7444 MeuSW4120 MeuSW4120 MeuSW7444 MeuSW7443 MeuSW4120 
Mgin MginNZ03 Mgin01TW Mgin01TW MginNZ03 Mgin01TW MginNZ03 MginNZ03 
Mgr MgrNZ67 MgrNZ67 MgrNZ55 MgrNZ55 MgrNZ55 MgrNZ01 MgrNZ02 
Mhe MheWA02 MheNZ02 MheNZ02 MheWA02 MheNZ02 MheWA02 MheMgr09 
Mlay MlayNZ10 MlayNZ08 MlayNZ08 MlayNZ10 MlayNZ10 MlayNZ10 MlaySAM9788 
Mmi MmiSW4968 MmiSW4972 MmiSW4968 MmiSW4972 MmiSW4968 MmiSW4968 MmiSW4968 
Mpi TMMCC75 TMMCC75 TMMCC75 TMMCC75 TMMCC75 TMMCC75 TMMCC75 
Mpe MpeLAM95654 MpeLAM95654 MpeLAM95654 MpeLAM95654 MpeLAM95654 MpeLAM95654 MpeLAM95654 
Mst MstSW4962 MstSW9491 MstSW9491 MstSW9491 MstSW9491 MstTSM30135 MstSW4962 
Zca ZcaNZ12 ZcaNZ12 ZcaNZ12 ZcaNZ12 ZcaNZ06 ZcaNZ06 ZcaNZ12 
Tsh TshNZ01 TshNZ01 TshNZ01 TshNZ01 TshNZ01 TshNZ02 TshNZ01 

Appendix 2

Maximum likelihood analyses of individual introns. Terminal nodes are labeled with species codes as per Figure 2. Bootstrap scores ≥50% are shown. Scores in bold font indicate strongly supported nodes. Asterisk in BGN tree highlights the only strongly supported node not observed in subsequent analyses of the concatenated (COMBO) data set.